摘要
传统激光脉冲下光电子谱识别方法受噪声影响,导致方法存在识别效率低、识别准确率低和抗噪声能力差的问题,为此提出基于人工智能技术的激光脉冲下光电子谱识别方法。在小波阈值去噪原理的基础上,通过三维融合策略对光电子谱图像进行去噪处理,避免噪声对光电子谱识别结果产生影响。采用人工智能技术中的深度卷积生成对抗网络,将去噪后的光电子谱输入卷积生成对抗网络中,输出光电子谱的识别结果。实验结果表明:所提方法的识别效率较高、识别准确率较高、抗噪声能力较强。
The traditional recognition method of photoelectron spectrum under laser pulse is affected by noise,which results in low recognition efficiency,low recognition accuracy and poor anti-noise ability.Therefore,a photoelectron spectrum recognition method under laser pulse based on artificial intelligence technology is proposed.On the basis of the wavelet threshold denoising principle,the photoelectron spectrum image is denoised by a three-dimensional fusion strategy to avoid the influence of noise on the photoelectron spectrum recognition result.The deep convolution generative adversarial network in artificial intelligence technology is used,and the denoised photoelectron spectrum is input into the convolution generative adversarial network,and the recognition result of the photoelectron spectrum is output.The experiment results show that the proposed method has high recognition efficiency,high recognition accuracy and strong anti-noise ability.
作者
李丽芳
班小强
LI Li-fang;BAN Xiao-qiang(School of Intelligent Manufacturing,Guangdong Nanfang Institute of Technology,Jiangmen 529000,Guangdong Province,China)
出处
《信息技术》
2022年第11期31-36,共6页
Information Technology
基金
2019年度全国职业教育科研规划课题(2019QZJ049)。
关键词
人工智能技术
小波阈值去噪
三维融合策略
深度卷积生成对抗网络
光电子谱识别
artificial intelligence technology
wavelet threshold denoising
three-dimensional fusion strategy
deep convolution generative adversarial networks
photoelectron spectrum recognition
